首页> 外文期刊>電子情報通信学会技術研究報告. パターン認識·メディア理解. Pattern Recognition and Media Understanding >Detection of Wrong Character Using Probability Transitional Patterns of Both-Direction N-gram Probabilities
【24h】

Detection of Wrong Character Using Probability Transitional Patterns of Both-Direction N-gram Probabilities

机译:使用双向N-gram概率的概率转换模式检测错误字符

获取原文
获取原文并翻译 | 示例
           

摘要

The method of detecting the wrong characters and correction is proposed to improve recognition precision in character recognition. We pay attention to a technique in which N-gram probabilities are used, and propose a method which detects misspelled characters by both direction N-gram probabilities. Context probability is also introduced to be incorporated with estimated probabilities transitional patterns in both direction N-gram probabilities and context probabilities at the misspelling point. Moreover, we discuss the causes why some misspelled characters are not detected in this method. In addition, we consider the effectiveness of the method in the experiment.
机译:为了提高字符识别的识别精度,提出了一种错误字符识别和纠正的方法。我们关注使用N-gram概率的技术,并提出一种通过两个方向N-gram概率检测拼写错误的字符的方法。还引入了上下文概率,将其与拼写错误点上方向N-gram概率和上下文概率的估计概率过渡模式合并。此外,我们讨论了此方法中未检测到某些拼写错误的字符的原因。此外,我们在实验中考虑了该方法的有效性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号